OC the cause of shorten CPU life?

[ScBlcksunshine]

This might sound kind of strange, but anyway, I have a 2.8c OC to 3.3 running pretty rock solid for the past year or so (24hrs under Prime95..etc) but recently I experienced frustrating random reboot or BSOD showing something like single bit ESP error detected from CPU. For a while, I was pulling my hair out trying to figure out what causes the random reboot, from replacing the MB, reformat the system, virus scan, memtest...etc still no luck, until I decided to get a new 3.0E P4 and the system is once again rock steady. The interesting part is when I downclock the orginal P4 back to 2.8 the random reboot would stop. My question is could this processor be going bad internally since it's been OC for a while? I always thought that if a processor is bad, it wouldn't even let you boot into windows. but in this case everything is ok if the processor isn't OC,( I can even run games like Farcry or Doom 3 for hrs without reboot) this led me to believe that maybe the OC damaged certain transistors within the core overtime to a point that it's now very sensitive to error when OCing. Does anyone else have a similar experience to that? or any comments on my speculation would be helpful
.

 

[xiaolanhua]

Well, yes it will...

But if you overclock with the safe voltage and keep it @ the right temperature, it will hardly shorten the CPU...

Also, it may take longer to die then you want to keep the CPU...

I had my 2.4C running @ 3.2ghz 24/7 for about 1.5 years now, and I have 0 CPU related problem...


Computer Maniac

 

[Vio1]

I noticed something like what you are experiencing, with my old 2.4c. When i first got it was rock solid at 3.4ghz. Prime stable for 24+ hours. After a few months of use the computer became a bit unstable... and so was running it at 3.3ghz... then a few months later it wasnt stable even at 3.2ghz. At that point I sold it, and bought a 2.8c.


BTW: WELCOME TO THE FORUM!

 

[ScBlcksunshine]

Well, I suspect that might be the problem since I know transistors can be sensitive to voltage changes and FSB . I was also reading the thread on SNDS on Northwood processor and noticed bunch of people are having the same problem too, now that I have a prescott 3.0 OCing to 3.5 hopefully it'll last me a bit longer. one question though, I heard the prescott is a bit more durable and have higher heat tolerance then the northwood, is there any truth to that?

 

[Vio1]

As far as I know, i dont think they made it anymore durable... intels CPUs i believe are supposed to work under 75C... the prescotts run warmer then the northwoods, and have heard that they run fairly well, even though they are warmer.

 

[Albuquerque]

So long as you never change the voltage, an overclocked processor should be no more prone to failure than a normal chip that's rated at the overclocked speed.

Example: 2.4C overclocked to 3.2ghz on stock voltage should be equal in lifespan to a 3.2C at stock speed.

Reason: The core of your processor is about the size of your pinky fingernail, but they're not build individually. Rather, many hundreds (thousands even) are build on 12" wafers, which are then laser-cut into individual pieces. Those individual pieces are packaged into what you're used to seeing as a real processor (PCB, pins, caps, heatspreader, etc) and then are speed tested, graded and binned.

The processors that test successfully at higher speeds have their MSR's programmed with a higher multplier and are sold as higher-speed parts. The processors that test successfully only at lower speeds will have their MSR's programmed with lower multipliers and will be sold as lower-speed parts. But they all came from the same wafer of silicon. Thus, a lower speed chip and a higher speed chip (both at the same voltages and temperatures) will have the same anticipated lifespan when running at the same speed.

What really chews up those transistors is voltage. I was posting to another thread earlier with some quick math: a ~6% increase in voltage can equate to as much as a 10% increase in electrical load in the processor. A 10% increase in voltage could mean as much as a 17-18% increase in load... That's where you can really burn up a processor.

 

[ScBlcksunshine]

That's good info to know, I got another question though, right now I have the 3.0E running good at 3.5 at 1.38 V pretty close to the stock voltage. Since it's such a small voltage increase do you think it'll shorten the life span dramtically? I just don't want to have to replace processor in a year or so

 

[*jj]

it should be fine. For some reason I thiink I remember hearing the MTBF in a processer at stock voltages will be something like 10 years (I could be off, but not by much). So even an oc'd processor running at stock voltages should last close to that. If you shorten its life it won't be by much..cooling is a large part of the equation. The better you can dissipate heat from the cpu the better.

nice tidbit of info Albuquerque!!!

 

[jszent]

I have overclocked for years with no problems with any of my processors........but I do think that the more you push it the more you take off the life of the chip.......
"its only money anyway"......... and ocing is an expensive hobby.

 

[Falkentyne]

This might sound kind of strange, but anyway, I have a 2.8c OC to 3.3 running pretty rock solid for the past year or so (24hrs under Prime95..etc) but recently I experienced frustrating random reboot or BSOD showing something like single bit ESP error detected from CPU. For a while, I was pulling my hair out trying to figure out what causes the random reboot, from replacing the MB, reformat the system, virus scan, memtest...etc still no luck, until I decided to get a new 3.0E P4 and the system is once again rock steady. The interesting part is when I downclock the orginal P4 back to 2.8 the random reboot would stop. My question is could this processor be going bad internally since it's been OC for a while? I always thought that if a processor is bad, it wouldn't even let you boot into windows. but in this case everything is ok if the processor isn't OC,( I can even run games like Farcry or Doom 3 for hrs without reboot) this led me to believe that maybe the OC damaged certain transistors within the core overtime to a point that it's now very sensitive to error when OCing. Does anyone else have a similar experience to that? or any comments on my speculation would be helpful
.

What voltage was the chip run at ? What was the default voltage? How much did you increase the vcore?

 

[Falkentyne]

I noticed something like what you are experiencing, with my old 2.4c. When i first got it was rock solid at 3.4ghz. Prime stable for 24+ hours. After a few months of use the computer became a bit unstable... and so was running it at 3.3ghz... then a few months later it wasnt stable even at 3.2ghz. At that point I sold it, and bought a 2.8c.


BTW: WELCOME TO THE FORUM!

Same question for you too:
What voltage was the chip run at ? What was the default voltage? How much did you increase the vcore?

 

[ScBlcksunshine]

What voltage was the chip run at ? What was the default voltage? How much did you increase the vcore?

I think the 2.8c default at around 1.5 and the most I got it up to is 1.625 and keep in mind this is on abit IC7 which is famous for undervolting. That's why it's weird cuz I know some people have higher voltage and yet their northwood will last more then a yr OC. In any case I am keeping it safe on the Prescott and only volting it at 1.38 at 3.53, hopefully that shouldn't affect the overal life expectancy too much. BTW is the prescott Strained Silicon more sensitive to voltage changes or about the same as Northwood?

 

[xTrEmEoVrClOcKr]

Same thing with my 2.6C. I had this puppy running @ 3.53Ghz Stable (Prime 95 12+ hours) and it only started giving me BSODs after about a month or 2. I had about 1.75V running through it and it didn't affect it. I had it running for 6 months @ 3.43Ghz @ 1.75. I had this chip for about 1.2 years and now its starting to degrade, along with my Asus . Im suprsied my 2.6C didn't get SNDS after that .

 

[jszent]

I think the 2.8c default at around 1.5 and the most I got it up to is 1.625 and keep in mind this is on abit IC7 which is famous for undervolting. That's why it's weird cuz I know some people have higher voltage and yet their northwood will last more then a yr OC. In any case I am keeping it safe on the Prescott and only volting it at 1.38 at 3.53, hopefully that shouldn't affect the overal life expectancy too much. BTW is the prescott Strained Silicon more sensitive to voltage changes or about the same as Northwood?

The C-7 is famous for undervolting along with the S-7 not over volting.........

 

[flunders]

If you overclocked a 3.2e to 3.9 with v1.55 but kept the temps down, is it still a serious problem? What is safe voltage for a prescott with good air cooling?

 

[uclajd]

I noticed something like what you are experiencing, with my old 2.4c. When i first got it was rock solid at 3.4ghz. Prime stable for 24+ hours. After a few months of use the computer became a bit unstable... and so was running it at 3.3ghz... then a few months later it wasnt stable even at 3.2ghz. At that point I sold it, and bought a 2.8c.

This exact thing happened to my 2.4c! Worked great at 3.4 for some time, now only 3.25.

 

[batboy]

That's good info to know, I got another question though, right now I have the 3.0E running good at 3.5 at 1.38 V pretty close to the stock voltage. Since it's such a small voltage increase do you think it'll shorten the life span dramtically? I just don't want to have to replace processor in a year or so

Intel sets the default vcore on Prescotts from something like 1.36v to 1.4v. I would say anything up to 1.4v is perfectly safe as long as you don't overheat it.

 

[batboy]

If you overclocked a 3.2e to 3.9 with v1.55 but kept the temps down, is it still a serious problem? What is safe voltage for a prescott with good air cooling?

Not sure we know exactly the safe vcore of a Prescott yet, but I generally recommend not exceeding 1.5v. If you go above 1.4v, then you probably should have good cooling.

 

[(HR)Titan]

recommend not exceeding 1.5v.

so the 1,5v is max limit

 

[batboy]

I think 1.5v is the safe limit for normal cooling with a Prescott (others might disagree). If you don't mind a little more risk and/or if you have super cooling, then you might use more vcore. But remember, this thread is about what shortens a CPU life and high vcore definitely reduces longevity.